top of page

Another doctor in our sHAI Group: Erkan Basar defended his PhD and answered the question how chatbots for promoting healthy lifestyles can be developed and evaluated


On November 25th, 2025, Erkan Basar successfully defended his dissertation at Radboud University Nijmegen. His research addresses the question: How can we develop and evaluate engaging long-term chatbots with adequate conversational skills for promoting healthy lifestyles?


Chatbots for Health Counselling

Erkan’s work provides insights from various perspectives that strengthen our understanding of developing and employing chatbots for health counseling. As traditional counseling is limited by resources and staff shortages, chatbots present a promising, scalable solution for immediate and accessible support. Today, people already turn to chatbots for health advice and support, but ensuring they are equipped to provide safe, ethical, and reliable guidance remains crucial. He found the optimal solution in the careful integration of the advantages offered by generative AI with the stability of more conventional methods. This combination enhances user engagement, along with enabling control and oversight over the generative AI’s application, while allowing coordination of the conversations based on responsible design principles and ethical guidelines.


Ethical Considerations

As part of a multidisciplinary collaborative effort, Erkan and his team emphasized that ethical considerations should guide every stage of chatbot-based health interventions, from development to practice, to ensure these tools serve as reliable facilitators of the counselling process. Furthermore, they proposed a theoretical framework outlining when chatbot interventions are appropriate and when they risk infringing on human autonomy, providing a step forward for an ethical application.



The Role of Generative AI

Erkan’s research reviewed the use of generative AI as health counselling chatbots and found that, while capable of producing more specific and engaging responses than pre-written human ones, they lack genuine understanding, emotional depth, and the ability to foster self-reflection. Further content analysis revealed that chatbots built only on generative AI are likely to be inadequate in helping individuals understand their personal desires and motivations, and will not account for human autonomy. Consequently, individuals should not use such chatbots as substitutes for human counsellors, and chatbot design should prioritize responsible, human-centered principles that respect users’ needs and values.


A Hybrid Chatbot Architecture

His studies also explored methods for developing chatbots that maintain long-term user engagement and provide appropriate health interventions by leveraging the strengths of generative AI in a controlled manner. He proposed a hybrid chatbot architecture that balances control and flexibility by dynamically adjusting generative AI’s involvement and supports structured, moderated, and responsive conversations, offering a versatile framework for future chatbot applications in sensitive domains like health. While the architecture is designed primarily to support responsible and engaging conversations, it also offers sustainability benefits by potentially reducing reliance on resource-intensive usage of generative AI.

Effectiveness of Chatbots

The research team further investigated the effectiveness of chatbots in health interventions. Their findings indicated that chatbots produced encouraging outcomes and positive user experiences during long-term interventions, highlighting their potential to offer consistent support and improve accessibility. However, due to ongoing technical, ethical, and practical challenges, chatbots should be considered as complementary tools to human counselors, ideally used alongside ongoing supervision and guidance from qualified professionals.


Read more about Erkan’s work here:


🚀 Congratulations, Erkan, on this big achievement and much needed research in the field of conversational technology development!

Comments


Commenting on this post isn't available anymore. Contact the site owner for more info.

© 2026 sHAI Group

bottom of page